back to blog

A Practical Perspective: AI for Decision-Makers

2024-08-02

opinion

Knowing where you can and should use AI to solve business problems relies on understanding what it does. While you probably don't need to know how it works under the hood, getting to grips with its raw value proposition is key.

If our severely simplified description haunts you and you're after something more technical, let us know! Or if it's bang on what you wanted to hear, we'd also love to hear from you. Any feedback or suggestions for future articles is always welcome!

At its core, a business is a collection of people making decisions to solve problems and achieve common goals. These goals vary from creating profits to supporting communities, advancing technology, or protecting the environment. In all cases, decision-making is fundamental.

Decisions arise from information, and our society has consistently moved towards increasing our available data. The information technology boom of the past three decades has accelerated this trend significantly. Now, we find ourselves inundated with information that far exceeds our cognitive capacity. We've reached—and likely surpassed—the limits of effective human decision-making [1].

AI presents an opportunity to enhance and offload decision-making processes. It allows us to automate human reasoning, addressing tasks previously exclusive to human cognition.

Consider AI as a custom operator in mathematics, akin to addition (+) or multiplication (×). It transforms inputs into outputs predictably, but with a crucial distinction: we can tailor how the system processes these inputs to produce desired outputs. This customisation is what makes AI particularly valuable in business contexts.

By exposing this custom operator to various scenarios, we train it to perform consistently when faced with new, similar situations. The primary costs associated with AI lie in collecting and processing these input scenarios.

Training an AI system requires a substantial number of examples. The complexity of the decision-making context directly impacts the cost of building the AI system. Simple scenarios result in lower costs, while complex contexts demand higher investments.

Consider these contrasting examples:

Dog Identification AI: Training involves feeding the system images of dogs. All necessary context is contained within a single image, making it relatively straightforward and cheap to acquire.

HR Decision-Making AI: Developing AI for hiring and firing decisions requires a much broader context. It demands comprehensive data including employee backgrounds, performance histories, and HR metrics, as well as an understanding of workplace dynamics. Capturing these nuanced factors is complex and potentially costly.

In essence, AI automates specific forms of human reasoning—teaching machines to complete tasks as a human would. Its value in business stems from its ability to enhance the decision-making processes that lie at a company's core.

As we progress through 2024, businesses across industries are discovering opportunities for AI implementation. The question for leaders becomes: How can you leverage AI to free up your teams' headspace, enabling them to focus on making more impactful decisions?

In our next post, we'll explore methods for calculating the ROI of AI in complex scenarios like HR decision-making, helping you navigate the cost-benefit analysis of AI implementation in your business.

[1] The scale of information overload is significant. In 2011, Americans consumed five times as much information daily as they did in 1986 (Levitin, 2014). On average, individuals make about 35,000 decisions per day (Sollisch, 2016).

This post is the first in a series designed to highlight the potential that AI possesses to unlock new business value. If it resonated, or you have any ideas for future content, get in touch at andrew@geppettoconsulting.co.uk.

Start learning today.